From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1758165Ab0EUG0L (ORCPT <rfc822;w@1wt.eu>);
	Fri, 21 May 2010 02:26:11 -0400
Received: from moutng.kundenserver.de ([212.227.126.187]:64255 "EHLO
	moutng.kundenserver.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1752213Ab0EUG0J (ORCPT
	<rfc822;linux-kernel@vger.kernel.org>);
	Fri, 21 May 2010 02:26:09 -0400
From: Arnd Bergmann <arnd@arndb.de>
To: Tyler Hicks <tyhicks@linux.vnet.ibm.com>
Subject: Re: [PATCH] vfs/eCryptfs: Handle ioctl calls with unlocked and compat functions
Date: Fri, 21 May 2010 08:25:32 +0200
User-Agent: KMail/1.12.2 (Linux/2.6.31-19-generic; KDE/4.3.2; x86_64; ; )
Cc: Frederic Weisbecker <fweisbec@gmail.com>, Ingo Molnar <mingo@elte.hu>,
       LKML <linux-kernel@vger.kernel.org>,
       Dustin Kirkland <kirkland@canonical.com>,
       Ecryptfs <ecryptfs-devel@lists.launchpad.net>,
       Thomas Gleixner <tglx@linutronix.de>, John Kacur <jkacur@redhat.com>,
       Alexander Viro <viro@zeniv.linux.org.uk>
References: <1274289855-10001-1-git-send-regression-fweisbec@gmail.com> <20100520233942.GA1086@ecryptfs> <20100520234223.GA1133@ecryptfs>
In-Reply-To: <20100520234223.GA1133@ecryptfs>
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Message-Id: <201005210825.32819.arnd@arndb.de>
X-Provags-ID: V01U2FsdGVkX1+FIUwTDdDpQpjyguYJpKeR8ZpB4SX07iTA0cF
 +3Fw1XXbci24JE87XnlXbldzSpk8Gn8zGxrKKK1Y3FN5M3GbbF
 hidj+aDpw6L4T27giWlKw==
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Friday 21 May 2010, Tyler Hicks wrote:
> Lower filesystems that only implement unlocked_ioctl aren't being
> passed ioctl calls because eCryptfs only checked for
> lower_file->f_op->ioctl and returned -ENOTTY if it was NULL.
> 
> eCryptfs shouldn't implement ioctl(), since it doesn't require the BKL.
> Instead, unlocked_ioctl() should be used and vfs_ioctl() can be called
> on the lower file since it handles locking, if necessary.  This requires
> vfs_ioctl() to be exported.

Calling vfs_ioctl doesn't help you at all here, you could simply call
the ->unlocked_ioctl function of the lower fs directly to do the same,
because ->ioctl will be gone soon.

You are howevers still missing a few calls that are done through do_vfs_ioctl
or file_ioctl.  To implement these, you need to add the file and super operations
that these call and forward the functions to the lower fs.

> +#ifdef CONFIG_COMPAT
> +static long
> +ecryptfs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
> +{
> +	long rc = -ENOTTY;
> +	struct file *lower_file = NULL;
> +
> +	if (ecryptfs_file_to_private(file))
> +		lower_file = ecryptfs_file_to_lower(file);
> +	if (lower_file && lower_file->f_op && lower_file->f_op->compat_ioctl)
> +		rc = lower_file->f_op->compat_ioctl(lower_file, cmd, arg);
> +	return rc;
> +}
> +#endif

You need to return -ENOIOCTLCMD here, not ENOTTY to cover the case where
the lower file system does not have a ->compat_ioctl function but has its
calls listed in fs/compat_ioctl.c.

	Arnd